Python 3.5 类型提示动态生成的实例属性

Python 3.5 type hinting dynamically generated instance attributes

我想为动态生成的对象属性添加 Python 3.5 类型提示,以便 IDE 正确地自动完成它们。 "dynamical" 是指在 class 创建期间或 __init__ 或任何其他方法中不存在该属性。

例如有没有办法通过评论或其他技巧添加这些?如果没有,我可以退而求其次添加虚拟 class 属性。

示例::

 class Request:
      """Example HTTP request object.

      We have `get_user()`  but we do not declare it anyhere.
      """

 ...


 # Pyramid's way of plugging in methods and properties to request, enabled addon capabilities for the framework
 # adds Request.user - done in different part of application lifecycle, not during class creation
 config.add_request_method(auth.get_user, 'user', reify=True)

我们的目标是使这项工作有效,以便 PyCharm 和其他 IDE 可以完成此属性。

  • 我子class编辑了真实的class

  • 我在 class

  • 中添加了伪 __type_hinting__ 方法
  • 我用这个class而不是真正的作为参数类型提示

    class Request(_Request):
        """
        HTTP request class.
        This is a Pyramid Request object augmented with type hinting information for Websauna-specific functionality.
        To know more about request read also
        * py:class:`pyramid.request.Request` documentation
        * py:class:`webob.request.Request` documentation
    
        Counterintuitively, this request is also available in non-HTTP applications like command line applications and timed tasks. 
        These applications do not get request URL from a front end HTTP webserver, but a faux request is constructed pointing to the website URL taken from ``websauna.site_url`` setting. 
        This is to allow similar design patterns and methodology to be applied in HTTP and non-HTTP applications.
    
        By setting variables in ``__type_hinting__()`` based on arguments types allows IDEs to infer type information when you hint your views as::
    
             from websauna.system.http import Request
    
             def hello_world(request: Request):
                 request.  # <-- When typing, here autocompletion kicks in.
    
        """
    
        def __type_hinting__(self, user: Optional[User], dbsession: Session, session: ISession, admin: Admin, registry: Registry):
            """
            A dummy helper function to tell IDEs about reify'ed variables.
            :param user: The logged in user. None if the visitor is anonymous.
            :param dbsession: Current active SQLAlchemy session
            :param session: Session data for anonymous and logged in users.
            :param admin: The default admin interface of the site. Note that the site can have several admin interfaces for different purposes.
            :param registry: Pyramid registry's. E.g. 
                :py:attr:`pyramid.registry.Registry.settings` for reading settings and :py:meth:`pyramid.registry.Registry.notify` for sending events.
            """
            self.user = user
            self.dbsession = dbsession
            self.session = session
            self.admin = admin
            self.registry = registry 
    

在 Python 3.6+ 中,您可以使用 class-level 类型提示 - 这些不会在 class 中生成属性。即

class Request(_Request):
    user: Optional[User]

这不会在 class 中创建属性,只会创建注释。

>>> Request.user
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
AttributeError: type object 'Request' has no attribute 'user'

>>> Request.__annotations__
{'user': typing.Union[foo.User, NoneType]}

在 Python 3.5 中可以创建一个 returns 一个 non-data 描述符的函数(即 没有 __set__);这将被实例属性覆盖,但它带有 一些最小的 运行时开销 - 将从 __dict__ 获取描述符并检查它是否定义了 __set__ 插槽- 即使是所有阅读。它可能看起来像

class Request(_Request):
    user = typed(User)

其中 typed 定义为

def typed(type: Type[T]) -> T:
    ... return a dummy non-data-descriptor...

这应该足以 PyCharm 正确推断类型。